WASTEWATER TREATMENT

(1)

DISSERTATIONES TECHNOLOGIAE CIRCUMIECTORUM UNIVERSITATIS TARTUENSIS

3

(2)

DISSERTATIONES TECHNOLOGIAE CIRCUMIECTORUM UNIVERSITATIS TARTUENSIS

3

PERFORMANCE

OF

WASTEWATER TREATMENT

WETLANDS

IN

ESTONIA

MÄRT ÖÖVEL

TARTU UNIVERSITY

P R E S S

(3)

Institute of Geography, Faculty of Biology and Geography, University of Tartu, Estonia.

The Faculty Council of Biology and Geography of the University of Tartu has, on 27 January 2006, accepted this dissertation to be defended for the degree of Doctor of Philosophy (in Environmental Technology).

Supervisor: Prof. Dr. Ülo Mander, Institute of Geography, University of Tartu

Opponent: Prof. Dr. Bjørn Kløve, Department of Process and Environ- mental Engineering, University of Oulu, Finland

Commencement: Scientific Council room in the University’s main building at Ülikooli 18; at 10:15 AM on 17 March 2006.

The publication of this dissertation has been funded by the Institute of Geography of the University of Tartu.

ISSN 1736–3349

ISBN 9949–11–234–6 (trükis) ISBN 9949–11–235–4 (PDF) Autoriõigus Märt Öövel, 2006 Tartu Ülikooli Kirjastus www.tyk.ee

Tellimus nr. 132

(4)

ORIGINAL PUBLICATIONS

Publication I

Mauring, T., Mander, Ü., Kuusemets, V. and Öövel, M. 2001. Efficiency of wastewater treatment wetlands in Estonia. In: Villacampa, Y., Brebbia, C.A.

and Uso, J.-L. (Eds) Ecosystems and Sustainable Development III, WIT Press, Southampton and Boston, pp. 479–491.

Publication II

Mander, Ü., Kuusemets, V., Öövel, M., Ihme, R., Sevola P. and Pieterse, A. 2000.

Experimentally constructed wetlands for wastewater treatment in Estonia.

Journal of Environmental Science and Health, Part A−Toxic/Haz. Subst. &

Environmental Eng., A35, 8, 1389–1401.

Publication III

Mander, Ü., Kuusemets, V., Öövel, M., Mauring, T., Ihme, R., Sevola P. and Pieterse, A. 2001. Wastewater purification efficiency in experimental treatment wetlands in Estonia. In: J. Vymazal (Ed.) Nutrient Transformations in Natural and Constructed Wetlands. Backhuys Publishers, Leiden, pp. 201–224.

Publication IV

Öövel, M., Tarajev, R., Kull, A. and Mander, Ü. 2005. Tertiary treatment of municipal wastewater in a floodplain peatland. In: De Conçeicao Cunha, M. and Brebbia, C.A. (Eds.) Water Resources Management III. WIT Transactions on Ecology and the Environment Vol. 80, WIT Press, Southampton, Boston, pp.

433–444.

Publication V

Öövel, M., Tooming, A., Mauring, T., and Mander, Ü. 200X. Schoolhouse wastewater purification in a LWA-filled hybrid constructed wetland in Estonia.

Ecological Engineering. (Submitted).

(6)

Author’s contribution

Publication I: The author is partly responsible for the fieldwork, data collection and analysis (about 50%), and for writing the manuscript (40%).

Publication II: The author is partly responsible for the fieldwork, data collection and analysis (50%), and for writing the manuscript (40%).

Publication III: The author is partly responsible for the fieldwork, data collection and analysis (50%), and for writing the manuscript (50%).

Publication IV: The author is responsible for writing the manuscript and data analysis (70%), and partly responsible for the fieldwork and data collection (40%).

Publication V: The author is responsible for writing the manuscript and data analysis (80%), and partly responsible for the fieldwork and data collection (40%).

(7)

ABSTRACT

In Estonia both natural, semi-natural and constructed wetlands are in use for wastewater treatment. In this thesis, the removal of organic material (BOD7), total nitrogen and total phosphorus in 8 treatment wetlands is analyzed. The studied wetlands were: a semi-natural wet meadow slope in Koopsi, an FSW channel with macrophytes in Sangla-Rakke, a VSSF sand/plant filter in Põlva, a hybrid wetland system in Kodijärve, an FSW wetland in Põltsamaa, a hybrid system in Kõo, a hybrid system in Paistu and the semi-natural floodplain of the Valgejõgi River. In the semi-natural wet meadow slope in Koopsi, 65% of BOD7, 67% of N and 80% of P was removed, and outflow values were in all cases below the standard limits. The outlet concentration of the heavily loaded bioditch in Sangla-Rakke varied – the average concentrations of BOD₇, N and P were 22, 5.4 and 1.4 mg l^–1 respectively. Apart from nitrogen, the VSSF sand/plant filter in Põlva showed satisfactory treatment efficiency: 82%, 36%

and 74% respectively for BOD₇, N and P. Similarly, nitrogen removal was insufficient in the hybrid wetland in Kodijärve, where the average outflow values for BOD₇, N and P were 13.4, 46.2 and 3.4 mg^–1 respectively. The FWS wetland in Põltsamaa did not work well, and outflow values varied greatly, in the case of BOD₇, N and P from 1.8 to 250, from 1.3 to 42 and from 1.6 to 9.7 mg l^–1 respectively. The hybrid system in Kõo works well in the case of BOD₇ and phosphorus, and purification efficiency for BOD₇, N and P was 87.9%, 65.5% and 72.3% respectively. The hybrid system in Paistu shows low output values of BOD₇ and P: 5.5, 19.2 and 0.4 mg l^–1 for BOD₇, N and P respectively.

The floodplain peatland for tertiary treatment shows low output values of nitrates and P: 0.2–1.8 and 1.5 mg l^–1 respectively. For domestic wastewater treatment in subsurface flow wetlands, the main problem was the insufficient removal of nitrogen. Nitrogen removal was higher in well aerated hybrid wetland systems. Semi-natural wetlands showed good performance in the treatment of secondary wastewater. The results show that hybrid CW systems consisting of subsurface flow filters can efficiently operate in conditions of very variable hydraulic load and cold winter conditions. Locally produced LWA as a filter material in CWs has shown good hydraulic conductivity and phosphorus sorption capacity. The Paistu CW can be considered one of the best systems in Estonia, with proper design and outstanding purification results.

(8)

1. INTRODUCTION

Natural wetlands have been used as convenient wastewater discharge sites for as long as sewage has been collected (Kadlec and Knight, 1996). Natural wetlands are still used for wastewater treatment, but at present the use of constructed wetlands is becoming more popular and effective around the world (Vymazal, 2001). Constructed wetland treatment systems are engineered systems that have been designed and constructed to utilize the natural processes involving wetland vegetation, soils and their associated microbial assemblages to assist in treating wastewater (Vymazal, 2001). They are designed to take advantage of many of the same processes that occur in natural wetlands, but do so within a more controlled environment (Vymazal et al., 1998). These systems have potential for the treatment of municipal wastewater, as well as from farms, landfills and some industrial areas. Because of their great volume, slow purification processes and heterogeneity, constructed wetland treatment systems are tolerant to changing hydraulic and nutrient loadings. This makes them more suitable than conventional treatment plants for the treatment of wastewater from individual houses, tourist resorts and other objects with variable wastewater flowrates and pollutant loadings. In addition, constructed wetlands have been used for the purification of wastewater from villages and small towns. As wetland treatment is an extensive technology with a great demand for area, the use of constructed wetlands for wastewater treatment is more suitable in sparsely populated areas. Constructed wetlands are able to remove suspended solids, organic matter, nitrogen, phosphorus, trace metals, bacteria and viruses, organic pollutants as well as other substances (chloride, sodium and potassium, sulphur, silicon) from wastewater. Various constructed wetland systems demonstrate great differences in purification efficiency. Constructed wetlands are usually designed and operated for wastewater treatment, but they can also have other functions.

There are about 800 small purification plants in Estonia. Since mostly small purification plants were constructed during the 1970s, they are now in bad shape. According to a study by Eesti Veevärk AS (2002), 40% of existing small purification plants work unsatisfactorily, and the main problem is the inadequate removal of nutrients. Because of that, as well as the high cost of new conventional purification plants, constructed wetlands are gaining popularity in Estonia (Kuusik, 1995). There are almost 30 wetland treatment systems in Estonia (Tooming, 2005), among which there are different examples of this technology. Despite the knowledge obtained from studies of existing constructed wetland in Estonia, as well as from other countries with cold climates, there is still a hesitant position concerning the use of constructed wetland treatment technology in Estonia (Tooming, 2005). For example, in a study for the determination of the best available technology for small purification plants,

(9)

constructed wetlands have been considered usable only for the treatment of wastewater from individual houses, especially during the warm period (Eesti Veevärk, 2002). Therefore it is important to gather and analyze data from various existing treatment wetlands over a longer period and develop suitable technologies for different conditions.

Objectives

The main objectives of this PhD dissertation are: 1) to analyze the performance of 8 existing Estonian treatment wetlands, and 2) to compare different types of constructed wetlands regarding their capacity for the removal of organic matter, nitrogen and phosphorus, 3) to highlight the most important positive and negative aspects of treatment wetlands in Estonian conditions.

(10)

2. OVERVIEW OF CONSTRUCTED WETLANDS FOR WASTEWATER TREATMENT

Constructed wetlands can be divided into different groups. The basic classification is based on the type of macrophytic growth, and further classification is based on the water flow regime (Vymazal, 2001) (Fig. 1). Different types of constructed wetlands can be combined with each other or with conventional treatment plants in order to take advantage of the best features of each wetland type.

Free-floating plants

Floating-leaved plants

Emergent plants*

Submerged plants

Surface flow (FWS)*

Sub-surface flow (SSF)*

Horizontal flow (HSSF)*

Vertical flow (VSSF)*

Hybrid systems*

Downflow*

Upflow

Constructed wetlands

Figure 1. Classification of constructed wetlands for wastewater treatment (Vymazal, 2001). * – types of constructed wetlands analyzed in current dissertation.

2.1. Constructed wetlands with free-floating plants

Free-floating plants have most of their photosynthetic parts above the surface of the water and their root systems below it. Typical plant species that have been used in large-scale applications are water hyacinth (Eichhornia crassipes) and duckweed species (Lemna, Spirodela, and Wolffiella) (Kadlec and Knight, 1996). Free-floating plants can be used for raw sewage as well as for primary or secondary treated effluents (Gumbricht, 1993).

(11)

The use in temperate climates of constructed wetlands with water hyacinth, one of the most productive plants in the world, is limited, because hyacinth needs high temperatures for growth (Vymazal et al., 1998). The major disad- vantages of duckweeds compared to waster hyacinth are their shallow root systems and sensitivity to wind, but their major advantage is their lower sensitivity to colder climates (U.S. EPA, 1998). Nevertheless, still treatment wetlands with duckweed in temperate climates can be used as seasonal (summer) wastewater treatment plants, since in winter they only work as anaerobic or facultative lagoons (Bonomo et al., 1997).

2.2. Constructed wetlands with floating-leaved macrophytes

Floating-leaved macrophytes include plant species that are rooted in the substrate, and their leaves float on the water surface. Nymphaea spp., Nuphar lutea and Nelumbo nucifera are typical representatives of this group. So far only a few systems have used this type of vegetation, and the use of constructed wetlands with floating-leaved species for wastewater treatment is considered questionable (Vymazal, 2001).

2.3. Constructed wetlands with submerged plants

The photosynthetic tissue of submerged aquatic plants is entirely submerged.

According to Gumbricht (1993), Cladophora spp, Enteromorpha spp, Pota- mogeton spp, Ceratophyllum spp, Myriophyllum spp, Elodea canadensis and E. nuttalli, Ulva lactuca and Egeria densa have been studied for wastewater treatment, but the use of submerged macrophytes for wastewater treatment is still in the experimental stage (Vymazal et al., 1998). The development of ephiphytic communities on the leaves of vascular plants may reduce photosynthesis in submersed macrophytes. Because of the shading of submersed macrophytes by algae and their sensitivity to anaerobic conditions, they have found their widest use as a tertiary treatment step, polishing the effluent or eutrophied natural waters (U.S. EPA, 1988; Gumbricht, 1993). Summing up different removal functions, the potential removal rate for submerged pond systems in temperate zones lies somewhere between 0,5–2 g N m^–2 d^–1 and 0,1–

0,3 g P m^–2 d^–1 (Gumbricht, 1993).

(12)

2.4. Constructed wetlands with emergent macrophytes

Constructed wetlands for wastewater treatment with emergent macrophytes can be constructed with many different designs. In general these can be categorized into two major groups according to their flow pattern: free water surface systems (FWS wetlands) and systems with sub-surface flows (SSF wetlands);

sub-surface flow wetlands can be further categorized into horizontal subsurface flow systems (HSSF or HF wetlands) and vertical sub-surface flow systems (VSSF wetlands) (Vymazal, 2001). Combinations of wetlands consisting of different wetland types are classified as hybrid systems (Vymazal et al., 1998).

2.4.1. Free water surface systems

A typical free water surface constructed wetland is a sequence of sealed shallow basins containing 20–30 cm of rooted soil, with a water depth of 20–40 cm, and dense emergent vegetation covering a significant part of the surface (Vymazal, 2001). Free water constructed wetland design variables include total area, the number, size, depth and shape of wetland cells, hydraulic retention time, vegetation types and coverage, inlet and outlet type and location, and internal flow parameters (U.S. EPA, 2000).

In free water surface wetlands, inflow water containing particulate and dissolving pollutants slows and spreads through a large area of shallow water and emergent vegetation. Settleable organics are rapidly removed in FWS wetlands by quiescent conditions, deposition and filtration (Kadlec and Knight, 1996). Attached and suspended microbial growth is responsible for soluble BOD (Vymazal, 2001). FSW wetlands typically have aerated zones and anoxic or anaerobic zones in deeper parts of ponds or in sediments. Major oxygen sources in FSW are surface aeration and photosynthesis carried out by phytoplankton, periphyton, and submerged plants (U.S. EPA, 2001). In FSW nitrogen removal may be achieved by plant uptake/harvesting, nitrification/

denitrification, volatilization and ion exchange, but it is most effectively removed by nitrification/denitrification (Vymazal, 2001; U.S. EPA, 2001).

According to Vymazal et al. (1998), FSW systems provide sustainable removal of phosphorus, but at relatively slow rates. Phosphorus removal in FSW systems occurs from adsorption, absorption, complexation and precipitation. However, precipitation with Al, Fe and Ca ions, as the major process in P removal, is limited by little contact between water column and the soil (Vymazal et al., 1998). Significant amounts of nutrients may be stored in sediments.

Based on the data from the literature, Vymazal (2001) has summarized the following design criteria and recommendations for FSW: pre-treatment – to at least the primary level; organic loading – <80 kg BOD₅ha^–1 d^–1; hydraulic loading – 0.7–5.0 cm d^–1; detention time: 5–15 days; aspect ratio (L:W) – 2:1 to

(13)

10:1; water depth – 0.4 m; bottom slope – 0.5%; soils – 20–30 cm to support the growth of emergent macrophytes, no special requirements for high hydraulic conductivity (local soil is used in many cases); vegetation – most commonly used species: in North America – Scirpus spp., Typha spp.; in Europe – Phragmites australis; harvest frequency – 3 to 5 years.

2.4.2. Horizontal subsurface flow systems

A horizontal subsurface flow constructed wetland is a constructed wetland consisting of a trench or bed underlain with an impermeable layer of clay or synthetic liner. The bed contains porous media that will support the growth of emergent vegetation, and wastewater flows slowly under the surface of the bed, more or less horizontally through the rhizosphere of the wetland plants. The most commonly used macrophytes are P. australis and Typha latifolia. The wastewater is treated by filtration, sorption and precipitation processes in the soil, and by microbiological degradation.

Organic compounds in the HSSF are degraded aerobically as well as anaerobically by bacteria attached to plants’ underground organs and media surface. The oxygen required for aerobic degradation is supplied directly from the atmosphere by the diffusion or oxygen leakage from macrophyte roots and rhizomes, although the oxygenation of the rhizosphere in HSSF constructed wetlands is insufficient (Vymazal, 2001). Nitrogen is removed in HSSF constructed wetlands by nitrification/denitrification, volatilization, adsorption and plant uptake, although the major removal mechanism is nitrification/

denitrification (Kadlec and Knight, 1996; Vymazal et al., 1998). As oxygenation of the rhizosphere in HSSF wetland systems is insufficient, incomplete nitrification is the major cause of limited nitrogen removal (Brix and Schierup, 1989; Vymazal, 2001). Phosphorus is primarily removed by adsorption and precipitation reactions with soil media (Jenssen et al., 1993, Kadlec and Knight, 1996). However, media used for HSSF wetlands (e.g. pea gravel, crushed stones) do not usually contain great quantities of Fe, Al and Ca, and therefore removal of phosphorus is generally low (Vymazal, 2001). This has led to investigations to find more efficient wetland media, such as the Light Weight Aggregates (LWA) or Light Expanded Clay Aggregates (LECA) (Zhu et al., 1997; Johansson, 1998; Jenssen and Krogstad, 2003, Adam et al., 2005) or shell sands (Søvik and Kløve, 2005). Several investigations have demonstrated that the assimilation of nutrients in plants in constructed wetlands play a minor role, usually less than 10% of nitrogen and less than 5% of phosphorus can be removed in constructed wetlands with harvesting (Geller et al., 1990; Mander et al., 2003; Toet et al., 2005).

According to data from various studies gathered by Vymazal (2001), the removal rate of nitrogen and phosphorus in HSSF wetland systems is 121–3817

(14)

g N m^–2y^–1 and 25–389 g P m^–2y^–1 respectively, in the case of nitrogen and phosphorus treatment efficiency was 11.4–76.4% and 4.8–65.0 % respectively.

Based on the data from the literature, Vymazal (2001) has summarized the following design criteria and recommendations for HSSF constructed wetlands: pretreatment: to at least the primary level; organic loading: <150 kg BOD₅ ha^–1d^–1 (usually <80 kg BOD₅ ha^–1d^–1); hydraulic loading: ST: < 5 cm d^–1, TT :< 20 cm d^–1; specific area: ST: approx. 5 m² PE^–1, TT: 1 m² PE^–1; detention time: >5 days; aspect ratio (L:W): 3:1 (could be <1:1); media: washed gravel, crushed stones (3–16 mm); hydraulic conductivity of media 10^–3–3.10^–3 m s^–1; media depth: 0.6–0.8 m (average); media porosity: 0.3–0.45; bottom slope:

1.0%; liner: HDPE, LDPE, PVC (thickness 0.5–1.00 mm); vegetation: most frequently used species: in Europe P. australis; in North America Scirpus spp., Typha spp. The most common difficulties experienced by wetland treatment systems have been related to maintaining partially aerated soil conditions (Kadlec and Knight, 1996). Authors have pointed out that when wetland systems are overloaded by oxygen-demanding constituents or are operated with excessive water depth, highly reduced conditions occur in sediments, resulting in plant stress and reduced removal efficiencies for BOD and ammonia nitrogen.

2.4.3. Systems with vertical subsurface flow

Constructed wetlands with vertical subsurface flow are quite similar to HSSF wetlands. The main differences between these systems lie in the feeding systems and in the direction of water flow in the filter media. VSSF wetland systems are intermittently fed with large batches, thus flooding the surface;

wastewater gradually percolates down the bed and is collected by a drainage network at the base (Vymazal, 2001). This kind of feeding leads to good oxygen transfer and hence makes aerobic purification processes possible. The pretreatment of inflow is usually needed to avoid clogging of the filter media.

The major treatment processes in a VSSF are the same as in an HSSF.

However, VSSF beds are far more aerobic than HSSF beds and are good for both nitrification and BOD removal, (Vymazal, 2001). Cooper (1999) re- commends 1 m² pe^–1 for BOD removal only, and 2 m² pe^–1for BOD removal and nitrification. On the other hand, VSSF beds do not provide much denitrification, and problems with denitrification may be solved using a two- stage plant (Vymazal et al., 1998). The removal of phosphorus is related to the choice of filter media.

According to Vymazal (2001), some VSSF wetlands have been constructed to use the upflow of wastewater (wastewater is brought to the filter bottom and passes the filter in an upwards direction).

(15)

2.4.4. Hybrid systems

Various types of wetlands can be combined in order to achieve higher treatment effect. However, hybrid systems most frequently comprise VSSF and HSSF systems arranged in a staged manner to promote nitrogen removal by creating conditions for both nitrification and denitrification (Vymazal, 2001). Apart from that, sub-surface flow wetlands can be combined with free-water surface wetlands.

(16)

3. MATERIAL AND METHODS

3.1. Description of selected treatment wetlands

Budgets of organic matter (BOD) and the total nitrogen and total phosphorus of the following systems are analyzed (Table 1; Fig. 1 and 2):

• A combined overflow-subsurface flow root-zone system on a Phalaris arundinacea-slope in Koopsi, Tartu County;

• An aquatic macrophyte channel (bioditch) in Sangla-Rakke, Tartu County;

• A vertical flow sand/plant filter in Põlva, Põlva County;

• Originally a two-chamber horizontal subsurface flow sand-plant filter with T. latifolia, Iris pseudacorus, and P. australis, now a hybrid system with a vertical flow filter, horizontal flow filter and phosphorus removal unit in Kodijärve, Tartu County;

• A cascade of 4 serpentine ponds with T. latifolia and P.australis, for secondary treatment of wastewater from the town of Põltsamaa, Jõgeva County;

• A hybrid wetland system consists of two vertical flow filters followed by a horizontal subsurface flow filter with P. australis in Paistu, Viljandi County;

• A hybrid wetland system consists of a two-bed vertical subsurface flow filter planted with P. australis, an HSSF filter planted with T. latifolia and P. australis, and two free water surface wetland beds planted with T.

latifolia in Kõo, Viljandi County;

• A floodplain on alluvial and peatland soil of the Valgejõgi River for secondary treatment of wastewater from the town of Tapa in Lääne-Viru County.

(17)

Table 1. Design parameters of selected treatment wetlands in Estonia.

Name Type Wastewater

type Year of

construction Area

(m²) Loading (PE¹) Koopsi Semi-natural wet

meadow slope with P.

arundinacea

Effluent from sedimentation pond

1989 2400 500

Sangla-

Rakke FSW channel with

helophytes (bioditch) Effluent from sedimentation pond

1989 140 190

Põlva VSSF sand/plant filter Effluent from

septic tank 1994 90 40

Kodijärve Hybrid system

(VSSF+HSSF) Effluent from

Põltsamaa FSW (cascade of

macrophyte ponds) Effluent from activated sludge plant

1997 12000 6670

Kõo Hybrid system (VSSF+HSSF+FSW)

Effluent from septic tank

2001 1200 300

Paistu Hybrid system

(VSSF+HSSF) Effluent from

Tapa Semi-natural floodplain

Effluent from activated sludge plant

2002 651 10000

1 – population equivalent, 1 PE = 60 g BOD7 d^–1.

(18)

Figure 1: Location of investigated systems in Estonia. A – the sand/plant filter in Põlva, B – the Phalaris-slope in Koopsi, C – the bioditch in Sangla-Rakke, D – the hybrid wetland system in Kodijärve, E – the Põltsamaa treatment wetland, F – the hybrid wetland system in Kõo, G – the hybrid wetland system in Paistu, H – seminatural experimental plot on floodplain peatland of Valgejõgi (Map source – Estonian Land Board).

3.2. Sampling and analysis

Water from inlet and outlet of treatment wetland systems in Põlva, Koopsi and Sangla-Rakke was sampled once a month from April 1989 to October 1995.

Water samples from the inlet and outlet of the Kodijärve system were taken once a month during the study period from January 1997 to April 2005. Since October 2002, after the establishment of the VSSF at Kodijärve, water samples were also taken from the outlet of the VSSF. The water discharge was measured automatically in the outlet using tipping buckets. In Põltsamaa the water samples were taken from the inlet of the first pond and from the outlets of all ponds, once a month from April 1997 to January 2001, and samples were also taken from April to June 2004 and from February to April 2005. The water discharge was measured from the inlet of the first pond and the outlet of the last pond. In Kõo the water samples were taken 8 times from October 2001 to February 2002 from the inflow of the system and the outflow of the VSSF, HSSF and FWS. In Paistu 18 series of water samples from October 2003 to October 2005 were taken from the inflow and outflow of the VSSF and HSSF.

(19)

A B C

D E F

G H

Figure 3: Design schemes of investigated systems. A – the sand/plant filter in Põlva, B – the Phalaris-slope in Koopsi, C – the bioditch in Sangla-Rakke, D – the hybrid wetland system in Kodijärve, E – the Põltsamaa treatment wetland, F – the hybrid wetland system in Kõo, G – the hybrid wetland system in Paistu, H – seminatural experimental plot on floodplain peatland of Valgejõgi (Mauring, et al, 2001, Publication I; Teiter, 2005; Öövel et al., 2005, Publication IV; Vohla et al., 2006; Öövel et al., 200X, Publication V). S1...S3 – sampling points.

(20)

In Tapa 5 series of water samples were taken from August to October 2002 from the inflow pipe and piezometers installed in the floodplain.

Water samples were analyzed for BOD7, SS, NH4+-N, NO2–-N, NO3–-N, total N, PO₄^3–-P and total P (all according to APHA, 1989) in the laboratories of Tartu Environmental Research Ltd. and Tartu Water Ltd.

In Kodijärve, soil samples were taken in September/October from 1997 to 2004 in both basins from a depth of 0–0.1 m, 0.3–0.4 m and 0.6–0.7 m and analyzed for N, P and organic matter (C) content at the Laboratory of Plant Biochemistry at the Estonian Agricultural University.

The purification efficiency (PE; %) of water quality indicators was calculated using the following equation (see Kadlec and Knight, 1996):

PE = (C_in–C_out)/C_in*100 (1)

where:

• Cin – average value of inflow concentration (mg L^–1);

• Cout – average value of outflow concentration (mg L^–1).

Mass removal (MR; g m^–2 d^–1) is calculated on the basis of the following equation (see Kadlec and Knight, 1996):

MR = {(C_in * Q_in) – (C_out*Q_Out)}/A (2) where:

• A – area of CW (m²);

• Qin and Qout – average values of water discharge in inflow and outflow (m³ d^–1);

• Cin and Cout – average concentrations in inflow and outflow (mg L^–1).

In the FWS system in Põltsamaa, these calculations are as follows:

PE = (CI in–CIV out)/CI in*100 (3)

where:

• CI in – average value of pond I inflow concentration (mg L^–1);

• C_IV out – average value of pond IV outflow concentration (mg L^–1).

MR = {(C_{I in} * Q_{I in}) – (C_IV out *Q_IV out)}/A (4) where:

• A – area of CW (m²);

• Q_{I in} and Q_{IV out}– average values of water discharge in inflow to pond I and outflow from pond IV (m³ d^–1);

• C_I_in and C_IV_out – average concentrations in inflow to pond I and outflow from pond IV (mg L^–1).

(21)

The removal of BOD7, total P, and total N in Kodijärve and Paistu was also described using an area-based first-order model (later called the k-C* model) (Kadlec and Knight, 1996; Kadlec, 2000):

ln[(C_o–C*)/(C_i-C*)] = –k/q (5) where:

• k= the area-based, first-order rate constant (m yr^–1);

• q= the hydraulic loading rate (m yr^–1);

• Co= the effluent concentration (g m^–3);

• C_i= the influent concentration (g m^–3);

• C*= the irreducible background wetland concentration (g m^–3).

Based on published data (Kadlec and Knight, 1996), the C* values of 1 mg l^–1 for BOD₇ and 1.5 mg l^–1 for total N were chosen. In the case of Kodijärve, the C* value of 0.9 mg l^–1 for BOD₇ was chosen, to use a lower value than the lowest outlet concentration. It is known that wetlands have very low natural total P background concentrations (Kadlec and Knight, 1996). The C* value for total P was assumed to be 0.03 mg l^–1.

The available data show that the effects of temperature on BOD and phosphorus removal are negligible in subsurface flow wetlands (Mander and Mauring, 1997; Wittgren and Maehlum, 1997; Noorvee et al., 2005b).

However, processes such as ammonification, nitrification and denitricication have been proven to be temperature-dependent. Therefore, rates of ammonia and total nitrogen will also be temperature-dependent (Kadlec, 2000). KT values for nitrogen reduction have to be converted to k20 values for purposes of comparison. The relation between kT and k20 is the Arrhenius equation:

kT= k20θ^T–20, (6)

where:

• k_T= the reaction rate coefficient at temperature T (^oC);

• k₂₀= the reaction rate coefficient at temperature 20 ^oC;

• θ= the temperature factor;

• T= temperature (^oC);

An estimate of the temperature factor of ammonia oxidation is θ=1.04 and for total N reduction θ=1.05 (Kadlec and Knight, 1996).

Outlet concentrations are compared with limit values set by the Water Act of Estonia, which are for purification plants with loading between 2000–9999 PE for BOD₇ and P 15 and 1.5 mg l^–1. For such a treatment plant, there are no limit values for N, and therefore in the case of N, the recommended limit value of 15 mg l^–1 is used. In the Water Act of Estonia, there are no outflow limit values for a treatment plant whose loading is less than 2000 PE. For small purification plants, outflow standard limits are set on a case by case basis. In this work, limit

(22)

values for smaller purification plants for BOD7, total N and total P are 20, 20 and 1.5 mg l^–1 respectively.

3.3. Statistical analysis

The statistical analysis of the data was performed using the programme Sta- tistica 6.0. The normality of variables was verified using the Kolmogorov- Smirnov, Lilliefors’ and Shapiro-Wilk’s tests. Apart from water and air temperature, water discharge and conductivity, the parameters’ distribution differed from normal, and hence non-parametric tests were performed. We used the Wilcoxon Matched Pairs Test and the Mann-Whitney U-Test to check the significance of differences between the inflow and outflow parameters. We also used Spearman Rank Correlation analysis to analyze the relationship between the water quality indicators. The level of significance of α = 0.05 was accepted in all cases.

For data interpolation in the semi-natural peatland plot in Tapa Kriging method is used. Kriging with linear variogram model is found to represent best the measured values. All sampling points were taken into account to interpolate values within the experimental area. Floating boundary conditions (i.e.

extrapolation) were allowed in calculation, while data lying outside the sampling area were truncated in the final data representation.

(23)

4. RESULTS AND DISCUSSION 4.1. Phalaris-slope in Koopsi

The seminatural wet meadow covered with P. arundinacea is used for the secondary treatment of dairy farm wastewater (flowrate 130 m³ d^–1). This semi- natural wetland is not insulated below to prevent groundwater pollution, and therefore it was possible to construct it without disturbing the natural plant cover.

Table 2. Inflow and outflow concentrations, purification efficiency and mass removal in the Phalaris-slope in Koopsi (average ± standard deviation). For non-normally distributed data average (av), min and max values are given. (Mander and Mauring, 1997).

Parameter BOD7 Total N Total P

Number of analysis (n) 23 23 23

Inflow (mg l^–1) av 17.0 16.0±5.8 4.1±2.7

min 4.0

max 70.0

Outflow (mg l^–1) av 4.0 5.0±2.5 0.7±0.4

min 1.0

max 18.5

Efficiency (%) 65±21 67±17 80±12

Mass removal (g m^–2 d^–1) 1.7±0.6 0.7±0.23 0.5±0.11 Because of primary purification in sedimentation ponds, inflow concentrations to the Phalaris-slope were quite low, especially in the case of organic matter.

Due to the sufficiently low areal loading and long detention time, the Phalaris- slope showed a high efficiency of total N and total P removal, and outflow values were significantly lower than effluent limit values. In the case of BOD, purification efficiency was relatively lower, although the average outlet value of BOD₇ was only 4 mg l^–1, which is comparable to the water quality of stream water (Table 2). The low output of N indicates that plant cover creates a pattern of aerobic and anaerobic zones where there are optimal conditions for both nitrification and denitrification. Effective P removal is apparently caused by the high iron content of the soil (Mauring, et al., 2001, Publication I). Nevertheless, part of the effluent is formed by overland flow, both nutrient removal and outlet concentrations have been relatively stabile.

(24)

4.2. Bioditch in Sangla-Rakke

In the case of the Sangla-Rakke wetland purification system, this is part of a drainage channel with intensive macrophyte growth, receiving farm wastewater (flowrate 125 m³ d^–1) that is purified in sedimentation ponds. Although purification took place in three aerobic and anaerobic ponds and a pond with T.

latifolia, influent parameters values were remarkably high. As a result of high input loading, the areal loading of 40 g BOD m^–2 d^–1 was considerably higher than suggested for FWS constructed wetlands. Although overloading, removal rates were highest among the studied wetlands (except for BOD7, which was the second highest) (Table 3). This supports the idea that until reasonable loading there is a significant positive correlation between input load and mass removal (Mander and Mauring, 1994; Mauring et al., 2001, Publication I).

As typical for an FSW, the removal efficiency of organic matter was satisfactory (81%), but due to high input loading the average concentrations in outflow slightly exceeded limit values. There were no big problems with N removal, because of relatively low input concentrations and satisfactory removal efficiency, concentrations in the outflow were well below standard limits. Although free water surface wetlands typically have a comparatively low P purification efficiency, P purification in the bioditch was satisfactory (69%), and concentrations in the outflow were low.

Table 3. Inflow and outflow concentrations, purification efficiency and mass removal in the bioditch in Sangla-Rakke (average ± standard deviation) (Mander and Mauring, 1997).

Inflow (mg l^–1) 136±112 17.8±10.8 5.2±3.6

Outflow (mg l^–1) 22±18 5.4±4.0 1.4±1.3

Efficiency (%) 81±9 66±12 69±19

Mass removal (g m^–2 d^–1) 3.5±2.7 2.7±2.0 1.6±1.2

4.3. Vertical subsurface flow sand/plant filter in Põlva

The vertical subsurface flow sand-plant filter in Põlva was built to treat wastewater from a group of individual houses. The area of the filter is 90 m², the upper part of the filter consists of sand and gravel, the lower part is made of soil, and the filter is planted with P. australis and T. latifolia. Sewage is pumped into the filter in intervals, and the wastewater flowrate was 2 m³ d^–1.

(25)

Despite the relatively high variation of BOD7 values in both the inlet (27–

460 mg O2 l^–1) and outlet (7–50 mg O2 l^–1) (Table 4), the average removal efficiency of BOD7 in the Põlva VSSF was satisfactory (82%). Nevertheless, average outflow concentration was slightly above the standard value. There were problems with N removal in the Põlva VSSF filter, as the average removal efficiency was only 36%, and inlet and outlet concentrations varied between 18–54 and 17–34 mg l^–1, respectively. Although VSSF filters are usually aerobic, the low removal of N in Põlva VSSF was caused by insufficient nitrification (Mauring et al., 2001, Publication I). The insufficient nitrification could be caused by weak vegetation development during the first two seasons. Iron release from the filter indicates anaerobic conditions in the deeper parts of the filter, which are suitable for denitrification. Occasional surface runoff may also result in less efficient purification.

Table 4. Inflow and outflow concentrations, purification efficiency, and mass removal in the vertical subsurface flow sand/plant filter in Põlva (average ± standard deviation).

For non-normally distributed data average (av), min and max values are given. (Mander and Mauring, 1997).

Inflow (mg l^–1) 173±114 40.5±10.6 10.9±4.2

Outflow (mg l^–1) av 28 24.8±5.9 2.6±2.0

min 6

max 160

Efficiency (%) 82±12 36±14 74±15

Mass removal (g m^–2 d^–1) 2.1±1.7 1.0±1.2 0.4±0.11 During the study period, the average inlet and outlet values of total P were 10.9 (from 6.2 to 22.0) and 2.6 (from 0.4 to 8.8) mg P l^–1 respectively. P removal was satisfactory, and the average removal efficiency was 74%, which is comparable to other studied subsurface flow wetlands. However, over a longer period there might be a saturation problem. The increasing Fe release from the sand filter indicates the increasing of anaerobic conditions and the reduction of P retention capacity.

4.4. The hybrid wetland system in Kodijärve

The horizontal subsurface flow (HSSF) stage of the Kodijärve hybrid constructed wetland system was constructed in 1996 to treat the wastewater of the hospital with average wastewater flowrate 4.2 m³d^–1. The double bed HSSF filter (312.5 m²) is filled with coarse iron-rich sand and covered predominantly with P. australis and Scirpus sylvaticus. In the summer of 2002 the system was

(26)

improved – between the outflow from the septic tank and the inflow to the wetland, a vertical subsurface flow (VSSF) filter (two intermittently loaded crushed limestone filled beds of a total area of 37.4 m²) was constructed. Also, an additional 10 m² phosphorus sedimentation filter was constructed at the outflow of the HSSF. In this paper, the purification processes of the VSSF and the HSSF are analyzed, observing inflow to the wetland system as outflow from the septic tank and outflow as the average outflow from both HSSF filters. The effect of the establishment of a VSSF and phosphorus sedimentation filter is analyzed in studies by Noorvee et al. (2005a) and Vohla et al. (2005).

Summary results of average inflow and outflow concentrations, purification efficiency and mass removal of the most important nutrients and organic matter are given in Table 5. Figure 4 shows the variation of the inlet and outlet concentrations.

In the case of Kodijärve, the concentrations of nutrients in the inflow to the wetland system are higher than is usual for domestic wastewater, caused by decreasing water consumption. Because of decreasing water consumption, the wetland system is operating on lower hydraulic loading than that for which it was originally designed. Table 5 indicates that the Kodijärve hybrid wetland system was satisfactorily efficient in terms of BOD₇ and P, which were 89%

and 75% respectively. On the other hand, nitrogen purification was less efficient, and the purification efficiency of nitrogen was 52%.

Table 5. Inflow and outflow concentrations, purification efficiency, and mass removal in the hybrid wetland system in Kodijärve (average±standard deviation). For non- normally distributed data average (av), min and max values are given.

Inflow (mg l^–1) 124.9±57.6 96.5±29.6 13.9±4.3

Outflow (mg l^–1) av 13.4 46.2±15.8 3.4±1.7

min 1.0

max 69.3

Efficiency (%) 89.0±12.8 52.1±19.0 75.2±18.7

Mass removal (g m^–2 d^–1) 1.6±1.5 1.2±1.0 0.18±0.16 The main problem encountered by the treatment plant was with nitrification, caused by anaerobic conditions dominant in deeper parts of HSSF beds (Mander et al., 2001, Publication III). Due to insufficient nitrification, the removal of NH₄-N was unsatisfactory (54%) and most of N in outflow was in form of NH₄-N. On the other hand, because of anaerobic conditions, denitrification was efficient and concentrations of NO₃-N in outflow were relatively low. To improve aeration in the wetland system, in the summer of 2002 a vertical flow wetland was built between the septic tank and the HSSF part, as the first stage

(27)

0 50 100 150 200 250 300

23.01.97 24.09.97 13.06.98 15.01.99 23.08.99 21.02.00 28.09.00 12.04.01 21.10.01 28.10.01 30.01.02 19.03.02 05.12.02 14.08.03 27.04.04 28.12.04

mgO l?¹

BOD7 Inflow BOD7 Outflow BOD7 limit

0 25 50 75 100 125 150 175 200

23.01.97 17.11.97 13.08.98 15.04.99 27.12.99 05.06.00 06.03.01 21.10.01 29.10.01 13.02.02 14.06.02 24.04.03 27.02.04 25.11.04

mg l¯¹

Total-N Inflow Total-N Outflow Total-N limit

0 5 10 15 20 25 30 35

23.01.97 17.11.97 13.08.98 15.04.99 27.12.99 05.06.00 06.03.01 21.10.01 29.10.01 13.02.02 14.06.02 24.04.03 27.02.04 25.11.04

mg l¯¹

Total-P Inflow Total-P Outflow Total-P limit

Figure 4. Variation of concentrations of organic material (after BOD7), N and P in inflow and outflow in the hybrid wetland system in Kodijärve.

(28)

of the hybrid wetland system. Due to the VSSF, aeration conditions improved in the wetland system. Hence there has been a significant improvement in the removal of organic material, and the mass removal rate of NH4-N and total nitrogen improved significantly, although the improvement of purification efficiency was not significant (Noorvee et al., 2005a). In the case of phosphorus, the mass removal rate significantly improved after the establishment of the VSSF, but unfortunately purification efficiency significantly decreased at the same time (Noorvee et al., 2005a). After the establishment of the VSSF, there have been problems with denitrification in the Kodijärve wetland system, reflected by increased average nitrate concentrations in outflow. Efficient denitrification is found to be retarded by improved aeration conditions in the HSSF, as well as by the low amount of organic material (Noorvee et al., 2005a).

During the investigation period, the outlet concentrations of phosphorus were slowly increasing, and at the same time the purification efficiency and annual phosphorus retention decreased. At the same time, there is constant Fe outwash from the HSSF, caused by anaerobic conditions in deeper parts of the HSSF, which decreases phosphorus sorption capacity. Thus the phosphorus retention capacity of the filter material in the HSSF wetland is apparently reaching its limit. To improve phosphorus removal, the additional 10 m² phosphorus sedimentation filter filled with the sediment from oil shale ash plateau was constructed at the outflow of the HSSF. The sedimentation filter showed satisfactory removal of phosphorus, but unfortunately there appeared saturation problems during relatively short time and filter needs some improvement (Vohla et al., 2005).

Although the average purification efficiency of phosphorus was satisfactory, the average outflow concentrations exceed limit values. Also, the outflow concentrations of nitrogen exceed the recommended level of 20 mg l^–1. In the case of organic matter, average outflow concentrations meet the limit value.

The average values of the area-based first-order rate-constant k for the BOD7, total N and total P of the hybrid CW throughout the whole study period were 15.8; 7.6; and 6.7 m yr^–1, respectively. In the case of the Kodijärve CW, the k values are lower than reported in the literature (Kadlec and Knight, 1996;

Kadlec, 2000). Lower k values are caused by low hydraulic load, as well as anaerobic conditions in the CW (Noorvee et al., 2005).

The N, P, and carbon contents in the filter material show a variable pattern in both in space and time (Mauring et al., 2001, Publication I). During the period from 1997 to 2003, the concentrations of N, P and C were generally increasing.

On the contrary, from 2003 to 2004 some decrement can be observed (Fig. 5).

(29)

0 100 200 300 400 500 600

1997 1998 1999 2000 2001 2002 2003 2004

mg/kg

0 2 4 6 8 10 12

g/kg

N P C

Figure 5. Change in average concentrations of N, P, and C in filter material of Kodijärve HSSF.

4.5. The free water surface treatment wetland system in Põltsamaa

The Põltsamaa free water surface constructed wetland is a cascade of 4 serpentine ponds designed for the secondary treatment of wastewater from the conventional treatment plant in Põltsamaa (about 5000 inhabitants). The total area of the CW reaches 1.2 ha, and the area of shallow ponds (average water depth 0.5 m, first pond 1.0 m) varies from 0.2–0.3 ha. The 2^nd and 3^rd pond were planted with T. latifolia and the 4^th pond with P. australis. Average inflow flowrate was 700 m³d^–1. This system was constructed in 1997.

Figure 6 shows the variation of the 1^st pond inlet and the 4^th pond outlet parameters in this system. The average inflow and outflow concentrations, purification efficiency and mass removal of the most important nutrients and organic matter is given in Table 6.

Table 6 indicates that there are great problems with purification processes in the Põltsamaa FSW. The problems are mostly caused by the malfunctioning of the Põltsamaa conventional treatment plant. Because of that, large variations in inflow concentrations and flowrates can be observed. Although the constructed

(30)

Figure 6. Inlet and outlet concentrations of organic material, N and P in the Põltsamaa free water surface wetland.

0 50 100 150 200 250 300 350

04.1998 06.1998 08.1998 10.1998 12.1998 02.1999 04.1999 06.1999 08.1999 10.1999 12.1999 04.2004 05.2004 02.2005 04.2005

mgO l¯¹

0 10 20 30 40 50 60 70 80 90

04.1998 06.1998 08.1998 10.1998 12.1998 02.1999 04.1999 06.1999 08.1999 10.1999 12.1999 04.2004 05.2004 02.2005 04.2005

mg l¯¹

0 2 4 6 8 10 12 14

04.1998 06.1998 08.1998 10.1998 12.1998 02.1999 04.1999 06.1999 08.1999 10.1999 12.1999 04.2004 05.2004 02.2005 04.2005

mg l¯¹

(31)

Table 6. Inflow and outflow concentrations, purification efficiency and mass removal in the free water surface treatment wetland in Põltsamaa (average±standard deviation). For non-normally distributed data average (av), min and max values are given.

Number of analysis 29 29 29

Inflow (mg l^–1) av 74.5 22.4±14.6 4.9±2.4

min 11.5

max 290.0

Outflow (mg l^–1) av 39.2 15.4±9.7 4.4±2.2

min 1.8

max 250.0

Efficiency (%) av 51.9 av 24.3 av –1.4

min -376.2 min -115.4 min -178.6 max 92.6 max 90.7 max 59.3 Mass removal (g m^–2 d^–1) av 2.5 av 0.5 av 0.1

min -4.7 min -1.0 min -0.3 max 12.9 max 4.9 max 0.5 wetlands are usually able to manage variable conditions, in Põltsamaa the variations in input loading seem to be too great. In the worst case the organic material loading reached 200 kg BOD5 ha^–1 d^–1, which significantly exceeds the recommended optimum organic loading rate of <80 kg BOD5 ha^–1 d^–1 for FSW (Vymazal, 2001). Also, in some cases there was a hydraulic overloading of the system. Due to the high and extremely variable input load, the BOD7, total N and total P values in the outlet of the FSW were high and extremely variable:

1.8–250; 1.3–42 and 1.6–9.7 mg l^–1 respectively. In the case of organic matter and P, the average outflow concentrations exceed limit values. The situation is the worst in the case of P, except for one case; outflow concentrations do not meet limit value and the average purification efficiency was negative. Low purification of P is typical for FSW wetlands. This is mainly caused by little contact between the water column and the soil, which limits precipitation with metals, as the major process in phosphorus removal. Removal of nitrogen was low, although average inflow concentrations were low and average outflow concentrations of nitrogen meet the recommended limit. Low nitrogen removal was in most cases caused by insufficient nitrification because there was no significant ammonia removal in this system. To improve purification in the Põltsamaa FWS wetland, it is important to guarantee the normal performance of the Põltsamaa conventional treatment plant, and it is also necessary to remove sediments from FWS ponds.

(32)

4.6. The hybrid wetland system in Kõo

The hybrid treatment wetland system at Kõo consists of a two-bed vertical subsurface flow (VSSF) filter (2×64 m², filled with 5–10 mm crushed limestone, planted with P. australis), an horizontal sub-surface flow (HSSF) filter (365 m², filled with 15–20 mm crushed limestone, planted with T. latifolia and P. australis), and two free water surface (FWS) wetland beds (3600 and 5500 m², planted with T. latifolia). The estimated inflow to the system is 40 m³d^–1. The wetland system was constructed in 2000 for the purification of the raw municipal wastewater generated by about 300 population equivalents.

Summary results of average inflow and outflow concentrations, purification efficiency and mass removal of the most important nutrients and organic matter are given in Table 7.

Table 7. Inflow and outflow concentrations, purification efficiency and mass removal in the hybrid wetland system in Kõo (average ± standard deviation). For non-normally distributed data average (av), min and max values are given.

Inflow (mg l^–1) 141±111.6 50.9±31.8 7.04±4.39

Outflow (mg l^–1) av 17.4 av 17.9 av 2.03 min 3.0 min 2.5 min 0.14 max 90.0 max 65.0 max 8.7 Efficiency (%) 87.9±10.9 65.5±24.4 72.3±24.6 Both inflow as well as outflow concentrations are variable in the Kõo hybrid wetland system. Limited data from the Kõo hybrid wetland system show satisfactory purification efficiency in the case of organic matter and phosphorus, although average effluent concentrations of phosphorus slightly exceed limit value. Nitrogen purification was less efficient, but the quality of outflow is satisfactory. In the case of nitrogen purification, there seems to be some problem with nitrification, reflected by relatively high NH₄-N concentrations in outflow (average NH₄-N outflow concentrations were 12.8 mg l^–1). This may be caused by inadequate aeration in VSSF caused by too high organic material loading. One can expect some improvement of purification, caused by the development of vegetation in the wetland system. No long-term conclusions, however, can be drawn on the basis of the limited data from the Kõo hybrid wetland system.

(33)

4.7. The hybrid wetland system in Paistu

The hybrid wetland system in Paistu consists of a two-chamber vertical subsurface flow (VSSF) filter (2×108 m²) and a 216 m² horizontal sub-surface flow (HSSF) filter bed. Both filters are filled with LWA (name of the local Estonian product: FIBO) of different sizes. The HSSF bed is planted with P. australis, whereas the VSSF beds are covered by topsoil and lawn. The treatment system was constructed in 2002, and it treats the wastewater of 140 people (120 students and 20 teachers and staff members, which for schoolhouses is calculated as 64 PE; Kuusik, 1995).

The summary results of average inflow and outflow concentrations, purification efficiency and mass removal of the most important nutrients and organic matter is given in the Table 8. Figure 7 shows the temporal variation of the inlet and outlet concentrations.

Table 8. Inflow and outflow concentrations, purification efficiency, and mass removal in the hybrid wetland system in Paistu (average±standard deviation). For non-normally distributed data average (av), min and max values are given.

Inflow (mg l^–1) 91.8±46.9 64.3±30.1 4.4±2.2

Outflow (mg l^–1) av 5.5 19.2±6.7 0.4±0.3

min 2.1

max 28.0

Efficiency (%) 90.8±13.1 62.8±21.6 88.6±11.3 Mass removal (g m^–2 d^–1) 1.53±1.28 0.48±0.42 0.06±0.04 Typically for schoolhouses, water discharge showed significant changes on both the diurnal and annual levels, being 7.4 m³d^–1 on average, and fluctuating from 0 (in night and in summer) to 17.7 m³d^–1. In conventional wastewater plants, such a change in hydraulic loading normally causes the collapse of purification processes (Wittgren and Maehlum, 1997). Nevertheless, in the Paistu hybrid CW system, no significant problems have been detected. Both the BOD₇ value and the concentrations of N and P increased significantly in the outflow from the HSSF, the respective average values were 5.5; 19.2, and 0.4 mg l^–1. A remarkable purification also occurred in the VSSF filter bed, although the purification of BOD was most significant. Average outflow concentrations met limit values for organic material, total-N and total-P.

In terms of purification efficiency and mass removal, the wetland system demonstrates outstanding results. The relatively high standard deviation values of mass removal are caused by changing hydraulic loading. Comparison of

(34)

Figure 7. Variation of concentrations of organic material, N and P in inflow and outflow of the hybrid wetland system in Paistu.

0 25 50 75 100 125 150 175 200

30.10.03 28.11.03 20.01.04 20.02.04 10.03.04 16.04.04 20.04.04 18.05.04 04.06.04 30.09.04 18.01.05 20.02.05 30.03.05 17.04.05 21.05.05 20.06.05 18.09.05 15.10.05

mgO l¯¹

0 15 30 45 60 75 90 105 120

30.10.03 28.11.03 20.01.04 20.02.04 10.03.04 16.04.04 20.04.04 18.05.04 04.06.04 30.09.04 18.01.05 20.02.05 30.03.05 17.04.05 21.05.05 20.06.05 18.09.05 15.10.05

mg l¯¹

0 12 34 56 7 8 109

30.10.03 28.11.03 20.01.04 20.02.04 10.03.04 16.04.04 20.04.04 18.05.04 04.06.04 30.09.04 18.01.05 20.02.05 30.03.05 17.04.05 21.05.05 20.06.05 18.09.05 15.10.05

mg l¯¹

WASTEWATER TREATMENT

PERFORMANCE

OF

WASTEWATER TREATMENT

WETLANDS

IN

ESTONIA

MÄRT ÖÖVEL

P R E S S

CONTENTS

ORIGINAL PUBLICATIONS

ABSTRACT

1. INTRODUCTION

2. OVERVIEW OF CONSTRUCTED WETLANDS FOR WASTEWATER TREATMENT

2.1. Constructed wetlands with free-floating plants

2.2. Constructed wetlands with floating-leaved macrophytes

2.3. Constructed wetlands with submerged plants

2.4. Constructed wetlands with emergent macrophytes

3. MATERIAL AND METHODS

3.1. Description of selected treatment wetlands

3.3. Statistical analysis

4. RESULTS AND DISCUSSION 4.1. Phalaris-slope in Koopsi

4.2. Bioditch in Sangla-Rakke

4.3. Vertical subsurface flow sand/plant filter in Põlva

4.4. The hybrid wetland system in Kodijärve

4.5. The free water surface treatment wetland system in Põltsamaa

4.6. The hybrid wetland system in Kõo

4.7. The hybrid wetland system in Paistu